Abstract :

This Project Aims To Develop A Deep Learningbased Intelligent Assistant, Nutribot, For Automated Food Image Classification And Nutritional Information Prediction. The Nutribot System Assists Users By Classifying Food Items Into Twelve Predefined Categories Apple Pie, Cannoli, Cheese Plate, Cheesecake, Chicken Wings, Chocolate Cake, Deviled Eggs, Donuts, French Fries, Frozen Yogurt, Ice Cream, And Macarons — And Instantly Providing Their Nutritional Breakdown Including Calories, Fat, Carbohydrates, And Protein Content Per 100 Grams. To Achieve This, Three Deep Learning Architectures Were Designed And Implemented: A Custom Convolutional Neural Network (CNN), SqueezeNet, And VGG16. These Models Were Trained, Validated, And Tested On A Curated Food Image Dataset. Transfer Learning Was Applied To VGG16 And SqueezeNet To Improve Classification Accuracy With Limited Data. The Custom CNN Was Built From Scratch To Demonstrate The Flexibility Of Tailored Architectures For Food Recognition Tasks. The System Is Deployed As A Flask-based Web Application With An Intuitive Interface Developed Using HTML, CSS, Bootstrap, And JavaScript. Users Can Register, Log In, Upload Food Images, Receive Instant Predictions With Confidence Scores, And Download Results. A MySQL Database Stores User Accounts, Prediction History, And Nutritional Reference Data. The Backend Supports Export Of Results In JSON And CSV Formats. Experimental Results Demonstrate That VGG16 Achieves The Highest Accuracy Of 91.4%, Followed By Custom CNN At 86.7% And SqueezeNet At 84.2% On The Test Set. The System Proves Efficient In Real-world Scenarios With Response Times Under 3 Seconds Per Prediction. This Nutribot Platform Promotes Healthier Dietary Habits By Enabling Precise Food Tracking And Nutritional Awareness.

Published:

19-6-2026

Issue:

Vol. 26 No. 6 (2026)

Page Nos:

1252-1257

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite